Management of power in server systems is important to avoid configurations that could cause cards or the system to fail. A series of power hungry cards can cause the power supply to exceed its rated capacity, forcing it to shutdown and bring down the entire server chassis. Existing products typically impose a maximum anticipated power limit that applies to all card slots, even if there is a range of useful applications that do not approve of this maximum. Extra power capacity is forced into the supply and hence added cost or undue current limitations on configurations typically exists.
It would be desirable for a server system to use a calculated power management method on a module-by-module basis, allowing the server system to be optimized for modules of differing power requirements.